Module 4: Historical GIS & Maps – Oh, say, say, say… Wait… We need to look at bias.[1]

This week started to scratch the surface regarding the current state of GIS (Geographical Information System) mapping and creating spatial visualizations using historic data. There’s a world of possibility that historians (and/or historical geographers) are starting to unlock and explore. Ian Gregory and Paul Ell pointedly note that, “GIS allows geographical data to be used in ways that are far more powerful than any other approach permits.”[2] GIS expanded this historian’s toolkit and introduced a way to “radically re-examine” the way that space is used and thought about. Written in 2000 by Loren Sibert, “Using GIS to Document, Visualize, and Interpret Tokyo’s Spatial History,” presents a detailed description of the time and effort it took to create their visualizations. At the time it took Sibert about two years to create the layers of geographic maps that included shorelines/rivers, administrative areas, population, and rail network along with “the attribute data, and linking the two through carefully designed coding systems.” [3]

Close to the same time that Sibert’s paper was published, I was an undergrad studying archaeology.  One of my earliest GIS experiences in the field was learning how to map the site we were working on in (nearly) real-time. At that point it was rather revolutionary to be able to overlay a map of the day’s fieldwork with a topographic map or aerial photograph. That previous work laid the foundations for thinking about future projects through the lens of the various geological aspects.  However, access to technology outside of an academic environment is often problematic.

I was very excited to learn about QGIS. However, I admit that I struggled with it. I followed what Dr. Otis was doing in class but bumbled through making it work for the 2014 election dataset. I could load the map. change colors, and even add labels, but couldn’t figure out how to visually manipulate the data. Eventually, I sought out some extra help on YouTube and found a beginner tutorial that was helpful for parts of the assignment and taught me how to use plugins to add a layer of Open Street Maps.[4]  

In the end, I created four versions of this map. One that reads as a typical election night map with precincts color coded to denote who won that specific area (Red Republican, Blue Democrat), along with another 3 maps that show the percentage of votes for the Democratic (Blue), Republican (Red), and Libertarian (Green) candidates across the provided precincts. However, I did not make maps for the two independent candidates. If I could have figured it out, I would have liked to create a map that included all the candidates with their percentage of the vote in gradients.  

• 

• 

• 

This also leads to talking about map bias and data gaps. In this specific instance we’re only seeing a part of the election night story for Congressional district 10. Only part of Fairfax County is included in our dataset, leaving out Clarke, Frederick, and Loudoun Counties, along with part of Prince William County. I also included the three candidate visuals to show how the data can be visually skewed. Although it could look like the Libertarian candidate made a good showing in Carson Franklin and a few other districts, in reality their best locations only had a max of 2.79% of the vote, compared with the republican max of 62% and the democrat max of 51%.  

Knowing the data gaps, I wouldn’t choose one of these to make an argument, but I did leave the Open Street Maps layer to give a greater context of place. This way, one can pan all the way to Winchester to see how far the district goes. Unfortunately, I couldn’t figure out how to make a new shapefile from the Open Street Map, otherwise I would have included that visualization as well.

Our class discussion and readings touched on how bias is implicit in maps and that there is no such thing as a neutral map.  I particularly enjoyed looking at the map of native lands, which is amazing in concept but also missing data due to colonialism.[5] I wanted to specifically look at the tribe who lived where I grew up. On the map they’re represented as the Appamattuck, However, the represented territory is incorrect and missing part of their land which was taken by the English Colonists and later became Bermuda Hundred. In this instance there is a bloody history, that has been suppressed partially by colonial mapmakers. Since maps are codified to be associated with authority and often go unquestioned, it’s important to take a step back and ask, “why was this map made?”; “what did the author stand to gain by its creation?”; and “what is it trying to convey?” Critically considering the map (and other historic data) may reveal more than what first meets the eye.

---

[1] Yeah, yeah, yeah…. I know, but how can an elder millennial resist adding lyrics to Maps (by the Yeah Yeah Yeahs ) while talking about maps?

[2] Ian Gregory and Paul Ell, “GIS and its role in historical research: an introduction”, in  Historical GIS: Technologies, Methodologies and Scholarship (Cambridge: Cambridge University Press, 2007), p.1

[3]  Loren Siebert, “Using GIS to Document, Visualize, and Interpret Tokyo’s Spatial History,” Social Science History 24, no. 3 (2000), p. 550.

[4] “An Absolute Beginner’s Guide to QGIS 3” Accessed September 19, 2021. https://www.youtube.com/watch?v=NHolzMgaqwE

[5] “Native Land” Accessed September 19, 2021. https://native-land.ca/